7 ms·

The only time in which you would need blockchain over a centralized DB is when everyone needs access to the DB, but no one trusts each other. Like if you needed

The only time in which you would need blockchain over a centralized DB is when everyone needs access to the DB, but no one trusts each other. Like if you needed to store healthcare information for multiple people which theoretically could be accessed by multiple competing hospitals or something. It's a lot of overhead to legitimize not just using a centralized DB for most applications.

nexuist 5y ago
For most applications, yes. But the "multiple people which theoretically could be accessed by multiple competing hospitals" is actually a real use case too. In many fields competitors could benefit from having access to the same curated, real time dataset. For example, a list of pharmacies or local insurance policy plans if you're a hospital. The problem is that none of the players in the game are incentivized to build this service because it requires someone to host the server on their infrastructure - and if you don't host it yourself, what happens if/when your partner decides to shut it down, or cut you out, or serve you false information? With a small blockchain each competitor can host their own version of the data and share new rows in an auditable, traceable, and permanent way (such that another user cannot go in and delete anything). It's easy to get lost in the hype, but peer to peer applications always serve valid use cases. A peer to peer database is no different even if it's enamored by cryptobros and scammers.
- XorNot 5y ago
  How would this solve anything? Competing for what? If a bunch of people are competing on say, price, then what's stopping them putting any number they want into the system, or similarly just rejecting every number anyone else puts into the system? They have no aligned interests in the blockchain, and outside of it the legal system already defines their responsibility (i.e. prices must be advertised accurately). "Blockchain" aka a Git repository is only useful when multiple trusted parties are trying to coordinate a central source of truth without inadvertently removing each others changes. At which point we can dispense with anything that sounds like cryptocurrency because it's all just signed commits.
madeofpalk 5y ago
Doesn't this only work if no one has the ability to take over 50% of the network?
- 2cb 5y ago
  Theoretically, but there are ways to protect against this. That's what proof-of-work and proof-of-stake are for instance. Certainly there is a huge financial incentive to run a 51% attack on Bitcoin or Ethereum but thus far no one has managed it.
KptMarchewa 5y ago
>The only time in which you would need blockchain over a centralized DB is when everyone needs access to the DB, but no one trusts each other. How is that not solved by a layer of permissions?
- wongarsu 5y ago
  If you have one trusted party that can run a database with permission system and audit trail, then you don't need a blockchain. But if no one trusts each other, but you still require everyone to have the ability to update some data, then a blockchain becomes useful
  - toomuchtodo 5y ago
    My multiple healthcare providers use OAuth and the electronic medical record system made by Epic to exchange my medical records, no blockchain required. Apple Health obtains access to all of the data at various providers through the same mechanism. At a higher level, if you’re willing to do business with someone, and you’re providing or accepting fiat, goods, or services, there is a baseline level of trust between parties and an understanding that any breakdowns in trust will be resolved by contract law and courts.
    - toomuchtodo 5y ago
      Replying to my own comment as the edit window has timed out. Apple has coincidently released functionality as part of iOS 15 for verifiable health records data. Note the use of digital signature crypto primitives. > Find out how you can securely request access to someone's verifiable health records and incorporate that data safely into your app. The Health app helps people download, view, and share their health records, including their COVID-19 immunization and test results — and iOS 15 brings support for the Smart Health Card, a verifiable health record that incorporates the FHIR health data standard. We'll show you how your app can go about requesting access to this record and how you can verify the signature of the file using CryptoKit and the issuer's public key. https://developer.apple.com/videos/play/wwdc2021/10089/ https://developer.apple.com/videos/play/wwdc2021/10089/
mytherin 5y ago
That use case is highly suspect. The reason sharing medical records is difficult is because medical records are sensitive information. Blockchain technology does not solve this problem, and in fact makes it worse. By the very nature of blockchain every single hospital would now need to store all of the medical information of everybody in the country. That means if a single hospital is compromised (which is inevitable) the medical history of every single person in the country will be leaked. At that point, you might as well just place everybody's medical records online for all to see, the end result will be very similar. Hospitals could encrypt this information inside the blockchain, but then they would need to contact each other for the keys, which defeats the entire point of the blockchain. At that point they might as well contact each other for the data, after all. Using blockchain technology also means you cannot abide by GDPR's right to forget. That means storing any personal information in a blockchain is a legal liability, at least in the EU, which limits its potential use cases even further.
- 2cb 5y ago
  > Using blockchain technology also means you cannot abide by GDPR's right to forget. That means storing any personal information in a blockchain is a legal liability, at least in the EU, which limits its potential use cases even further. Yes indeed, blockchains aren't suitable for storing personal data. No public decentralised network is. IPFS wouldn't be either for instance. By nature such networks assume no one peer can be trusted. Not the sort of thing you want to trust with confidential information. This openness is often sold as a feature, for example running your supply chain through a blockchain means there is a publicly auditable ledger of every step, so a company claiming to have an ethical supply chain could theoretically point to that blockchain as proof. This is potentially interesting imo. Most enterprise blockchain solutions I've seen are hybrid ones. A business can store personal data in a regular database and have non-sensitive data on a blockchain (as per the example above). The advantage of this (outside of the above use case) being that you can run SQL queries on large data sets much faster if that data is stored on a distributed ledger. (Note: this is what the companies selling those solutions claim. I don't have first hand experience with blockchains in an enterprise context so I couldn't tell you if this is true - I imagine it'd depend on the database and blockchain in question.) Final point though, these aren't necessarily proprietary blockchain solutions as such. They're more like SDKs businesses can use to build their own blockchains, code their own smart contracts, etc. Whether this is actually useful for enterprise... I honestly couldn't tell you. I do think the supply chain stuff is interesting. I also think using it as a system to detect counterfeit items is another good use case for businesses. In the past, companies have created apps where you can scan a QR code and it confirms the legitimacy of the product, but counterfeiters just made QR codes that tricked the app. If each unit is tracked on a public blockchain, it should be possible to verify legitimacy with near 100% certainty.

nexuist 5y ago

For most applications, yes. But the "multiple people which theoretically could be accessed by multiple competing hospitals" is actually a real use case too. In many fields competitors could benefit from having access to the same curated, real time dataset. For example, a list of pharmacies or local insurance policy plans if you're a hospital. The problem is that none of the players in the game are incentivized to build this service because it requires someone to host the server on their infrastructure - and if you don't host it yourself, what happens if/when your partner decides to shut it down, or cut you out, or serve you false information? With a small blockchain each competitor can host their own version of the data and share new rows in an auditable, traceable, and permanent way (such that another user cannot go in and delete anything). It's easy to get lost in the hype, but peer to peer applications always serve valid use cases. A peer to peer database is no different even if it's enamored by cryptobros and scammers.

XorNot 5y ago
How would this solve anything? Competing for what? If a bunch of people are competing on say, price, then what's stopping them putting any number they want into the system, or similarly just rejecting every number anyone else puts into the system? They have no aligned interests in the blockchain, and outside of it the legal system already defines their responsibility (i.e. prices must be advertised accurately). "Blockchain" aka a Git repository is only useful when multiple trusted parties are trying to coordinate a central source of truth without inadvertently removing each others changes. At which point we can dispense with anything that sounds like cryptocurrency because it's all just signed commits.

madeofpalk 5y ago

Doesn't this only work if no one has the ability to take over 50% of the network?

2cb 5y ago
Theoretically, but there are ways to protect against this. That's what proof-of-work and proof-of-stake are for instance. Certainly there is a huge financial incentive to run a 51% attack on Bitcoin or Ethereum but thus far no one has managed it.

KptMarchewa 5y ago

>The only time in which you would need blockchain over a centralized DB is when everyone needs access to the DB, but no one trusts each other. How is that not solved by a layer of permissions?

wongarsu 5y ago
If you have one trusted party that can run a database with permission system and audit trail, then you don't need a blockchain. But if no one trusts each other, but you still require everyone to have the ability to update some data, then a blockchain becomes useful
- toomuchtodo 5y ago
  My multiple healthcare providers use OAuth and the electronic medical record system made by Epic to exchange my medical records, no blockchain required. Apple Health obtains access to all of the data at various providers through the same mechanism. At a higher level, if you’re willing to do business with someone, and you’re providing or accepting fiat, goods, or services, there is a baseline level of trust between parties and an understanding that any breakdowns in trust will be resolved by contract law and courts.
  - toomuchtodo 5y ago
    Replying to my own comment as the edit window has timed out. Apple has coincidently released functionality as part of iOS 15 for verifiable health records data. Note the use of digital signature crypto primitives. > Find out how you can securely request access to someone's verifiable health records and incorporate that data safely into your app. The Health app helps people download, view, and share their health records, including their COVID-19 immunization and test results — and iOS 15 brings support for the Smart Health Card, a verifiable health record that incorporates the FHIR health data standard. We'll show you how your app can go about requesting access to this record and how you can verify the signature of the file using CryptoKit and the issuer's public key. https://developer.apple.com/videos/play/wwdc2021/10089/ https://developer.apple.com/videos/play/wwdc2021/10089/

mytherin 5y ago

That use case is highly suspect. The reason sharing medical records is difficult is because medical records are sensitive information. Blockchain technology does not solve this problem, and in fact makes it worse. By the very nature of blockchain every single hospital would now need to store all of the medical information of everybody in the country. That means if a single hospital is compromised (which is inevitable) the medical history of every single person in the country will be leaked. At that point, you might as well just place everybody's medical records online for all to see, the end result will be very similar. Hospitals could encrypt this information inside the blockchain, but then they would need to contact each other for the keys, which defeats the entire point of the blockchain. At that point they might as well contact each other for the data, after all. Using blockchain technology also means you cannot abide by GDPR's right to forget. That means storing any personal information in a blockchain is a legal liability, at least in the EU, which limits its potential use cases even further.

2cb 5y ago
> Using blockchain technology also means you cannot abide by GDPR's right to forget. That means storing any personal information in a blockchain is a legal liability, at least in the EU, which limits its potential use cases even further. Yes indeed, blockchains aren't suitable for storing personal data. No public decentralised network is. IPFS wouldn't be either for instance. By nature such networks assume no one peer can be trusted. Not the sort of thing you want to trust with confidential information. This openness is often sold as a feature, for example running your supply chain through a blockchain means there is a publicly auditable ledger of every step, so a company claiming to have an ethical supply chain could theoretically point to that blockchain as proof. This is potentially interesting imo. Most enterprise blockchain solutions I've seen are hybrid ones. A business can store personal data in a regular database and have non-sensitive data on a blockchain (as per the example above). The advantage of this (outside of the above use case) being that you can run SQL queries on large data sets much faster if that data is stored on a distributed ledger. (Note: this is what the companies selling those solutions claim. I don't have first hand experience with blockchains in an enterprise context so I couldn't tell you if this is true - I imagine it'd depend on the database and blockchain in question.) Final point though, these aren't necessarily proprietary blockchain solutions as such. They're more like SDKs businesses can use to build their own blockchains, code their own smart contracts, etc. Whether this is actually useful for enterprise... I honestly couldn't tell you. I do think the supply chain stuff is interesting. I also think using it as a system to detect counterfeit items is another good use case for businesses. In the past, companies have created apps where you can scan a QR code and it confirms the legitimacy of the product, but counterfeiters just made QR codes that tricked the app. If each unit is tracked on a public blockchain, it should be possible to verify legitimacy with near 100% certainty.